Add Progress Bar

[work4cs]

Could we add progress bar (such as tqdm) when we add a list of images to TreeOfLifeClassifier or CustomLabelsClassifier?

[johnbradley]

Thanks for the issue @work4cs.

When you pass a list of images to the predict() function these images are processed as a group. If you are using a "cuda" or "mps" device the images will be loaded onto the GPU and processed in a single encode_image operation. Typically this method will be faster than processing the images one at a time.

@work4cs Which device are you using "cpu", "cuda", or "mps"(default is "cpu")? How many images are you passing to the predict() function?

[work4cs]

Thanks for the issue @work4cs.

When you pass a list of images to the predict() function these images are processed as a group. If you are using a "cuda" or "mps" device the images will be loaded onto the GPU and processed in a single encode_image operation. Typically this method will be faster than processing the images one at a time.

@work4cs Which device are you using "cpu", "cuda", or "mps"(default is "cpu")? How many images are you passing to the predict() function?

Oh it seems like we are dealing with a single batch now. If a batch of images can fit in GPU or CPU, then we don't need to monitor the time and progress since it can finish in the blink of an eye. Would you like to extend it to multiple batch with a batch_size parameter passing to BaseClassifier for the case when there are too many images in the list? Thank you!

[johnbradley]

I think a batch_size parameter might be a little more convenient on the predict() function because the constructor for the classifiers takes a little while to load the model and supporting files.

[work4cs]

I think a batch_size parameter might be a little more convenient on the predict() function because the constructor for the classifiers takes a little while to load the model and supporting files.

Sounds good!